Knitted Fabric Pattern Design Based on Three-Dimensional Data Graphic Generation Algorithms

Abstract

In the field of contemporary textile pattern design, floral patterns are widely favored due to their diversity and aesthetic appeal. However, due to the wide variety of flowers and their complex morphological structures, accurately presenting their characteristics and aesthetic appeal through computer-aided design techniques, such as 3D modeling, poses certain challenges. Based on this, the study proposes a graphic editing algorithm that combines 3D data with a directional bounding box method. On this basis, a textile pattern design system is established. The results show that the proposed algorithm performs exceptionally well. Experimental results on the Oxford 102 Flower Dataset indicate that the proposed algorithm achieves a mean squared error of 0.25, a root mean squared error of 0.001, and a mean absolute error of 0.0017. In addition, the method for extracting contour sampling points based on the front and side angles of flowers maintains flower contour integrity over 85% throughout the experimental process, reaching 100% at most. These findings suggest that the proposed method can comprehensively capture flower-contour information and generate two-dimensional design drawings with high accuracy, thereby improving the efficiency of textile pattern design.